A method for defining targets in hyperspectral image data of a scene, has the steps: (1) identifying a target within the scene; (2) fitting a plurality of conjoined hyperplanes about the target in the scene; and (3) fitting a shape to the identified target by shrinkwrapping the hyperplanes about the target by translating and rotating the hyperplanes until a specified fit parameter is satisfied.
Legal claims defining the scope of protection, as filed with the USPTO.
1. A method for defining targets in hyperspectral image data of a scene, comprising the steps: identifying a target within said scene; fitting a plurality of conjoined hyperplanes about said target in said scene; and fitting a shape to said identified target by shrinkwrapping said hyperplanes about said target by translating and rotating said hyperplanes until a specified fit parameter is satisfied.
2. A method as in claim 1 , further comprising the step of spatial filtering to produce further information about the target.
3. A method as in claim 1 , further comprising the steps of: obtaining pixels representing the target and other information present in a locale of the target; partitioning the pixels into objects; performing a two-dimensional principle component analysis on the objects to determine a width and length of each of said objects; obtaining the overall size of each of said objects measured in total number of lit pixels; and selectively eliminating some of said objects, said elimination being based on actual size, pixel dimensions, or both said actual size and pixel dimensions, as selective elimination criteria.
Cooperative Patent Classification codes for this invention. Click any code to explore related patents in that topic.
November 3, 1999
August 3, 2004
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.